JOHN R. GIBSON, Senior Circuit Judge.

This case involves a dispute between competing breeders of corn seed. The district court awarded $46,703,230.00 to Pioneer Hi-Bred International based on Holden Foundation Seeds, Inc.'s misappropriation of the genetic make-up of certain seed corn. Holden contests the district court's liability determination and damage award on numerous grounds. Pioneer cross-appeals from the district court's denial of prejudgment interest. We affirm.

The bench trial in this case consumed over 37 days, spread over 14 weeks. The massive record consists of over 9,000 pages of transcript and 1,100 exhibits. Thus, our recitation of the facts is, of necessity, somewhat abbreviated.

The sale of hybrid seed corn is a multi-billion dollar industry. Pioneer is a vertically integrated seed corn company. It conducts a breeding program, develops parent seed, and produces hybrid seed corn for the retail market. The superior performance of its products, obtained in part through millions of dollars spent annually on corn research and development, has enabled it to gain a sizeable portion of the retail seed corn market. Its marketed seed corns include Pioneer hybrids 3780 (Pioneer's leader in sales for several years) and 3541 — which share a common parent, designated H3H.

Holden indirectly competes with Pioneer. Holden is a foundation seed company that develops inbred parent seed lines and sells these lines to its customers, also seed corn companies, which use them to produce hybrid seed in competition with Pioneer. During the 1980s, Holden's LH38, LH39, and LH40 ("LH38-39-40") were among its most popular parent lines.

Pioneer sued Holden on a number of legal theories, Claiming that Holden developed LH38-39-40 from misappropriated Pioneer H3H or H43SZ7 — protected trade secrets of Pioneer ("H3H/H43SZ7"). During discovery, the district court decided, at Pioneer's request, that the nature of this dispute required that Holden "freeze in" a particular story regarding the development of LH38-39-40. The court did so to prevent Holden from altering its story to conform to the scientific evidence eventually introduced. Holden claimed that although LH38-39-40 demonstrated some similarity to Pioneer's seed lines, Holden developed LH38-39-40 from an internal line, designated L120. The district court held that the genetic make-up of H3H/H43SZ7 is a protected trade secret of Pioneer. The court determined that although LH38-39-40 differed in some respects from Pioneer's seed, the lines had nonetheless been derived from misappropriated Pioneer material. The district court based its decision on expert testimony that LH38-39-40 derived from H3H/H43SZ7 and Holden's inability to offer adequate evidence of its "L120 story." The court also ruled in Pioneer's favor on its Lanham Act, interference with business advantage, unjust enrichment, and conversion claims. The district court awarded over 46 million dollars in damages to Pioneer. Holden appeals.

I.

Much of Holden's argument before this court attacks Pioneer's scientific evidence as inadequate to support the district court's judgment. A threshold issue, however, is Holden's assertion that the district court erred in admitting the scientific evidence relating to electrophoresis, liquid chromatography, and growout testing.

A.

The evidence from the electrophoresis, liquid chromatography, and growout testing was central to the district court's decision. The judge appointed an expert, Dr. Charles F. Lewis, to assist the court, and both sides relied heavily on expert testimony to support their positions. Nonetheless, Holden argues that the scientific evidence from these three tests failed to meet the "general acceptance" requirement of Frye v. United States, 293 F. 1013, 1014 (D.C. Cir. 1923), and this court's application of Frye in United States v. Two Bulls, 918 F.2d 56, 60-61 (8th Cir. 1990).

After briefing in this case, the Supreme Court decided Daubert v. Merrell Dow Pharmaceuticals, Inc., ___ U.S. ___, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), which held the "austere" Frye standard "should not be applied in federal trials." Id. ___ U.S. at ___, 113 S.Ct. at 2794; see United States v. Martinez, 3 F.3d 1191, 1196 (8th Cir. 1993), cert. denied, ___ U.S. ___, 114 S.Ct. 734, 126 L.Ed.2d 697 (1994). Rule 702 of the Federal Rules of Evidence, the Court stated, displaced Frye and the "general acceptance" standard. Daubert, ___ U.S. at ___, 113 S.Ct. at 2794. Thus, the focal point of Holden's admissibility argument before this court has been removed. Moreover, we are convinced that the district court considered the factors discussed in Daubert, and committed no error in admitting the evidence.

Before admitting scientific expert testimony, the trial court must conclude, pursuant to the Federal Rules of Evidence, that the proposed testimony constitutes: "(1) scientific knowledge that (2) will assist the trier of fact to understand or determine a fact in issue." Id. ___ U.S. at ___, 113 S.Ct. at 2796; Martinez, 3 F.3d at 1196. This standard is "a flexible one" whose "overarching subject is the scientific validity — and thus the evidentiary relevance and reliability — of the principles that underlie a proposed submission." Daubert, ___ U.S. at ___, 113 S.Ct. at 2797. The court's "focus, of course, must be solely on principles and methodology, not on the conclusions that they generate." Id.

The three tests rested on basic principles of genetics and horticulture, which the district court considered in some detail. Although the court did not expressly analyze these issues under the as-yet-unannounced Daubert formulation, its analysis convinces us that it did consider the experts' testimony to be scientific knowledge that would assist in determining facts in issue. Indeed, the court based its liability finding in large part on the ability of these tests to show Holden's possession of Pioneer's protected material. Pioneer Hi-Bred Int's v. Holden Found. Seeds, Inc., at [*]13-30, No. 81-60-E, slip op. at 31-68, 1987 WL 341211 (S.D.Iowa Oct. 30, 1987). The district court acknowledged that the tests were widely used in the industry, conducted by qualified experts, and considered helpful enough to be relied upon by both parties in their attempts to support their respective positions. Thus, the district court's opinion shows that it considered and applied many of the factors enumerated in Daubert.

Moreover, Holden's admissibility arguments to this court primarily focused on how the district court used the scientific evidence, not on its decision to admit the evidence. Holden explicitly concedes that "the scientific tests were admissible to show differences. . . ." In light of this concession, the analysis made by the district court, and the less rigid standard enunciated in Daubert, we cannot say the court erred in admitting the testimony. Holden also suggests that the tests were inadmissible because they could not conclusively show derivation. This inability, however, provides no bar to the tests' admission. " Daubert requires only scientific validity for admissibility, not scientific precision." United States v. Bonds, 12 F.3d 540, 558 (6th Cir. 1993).

This aspect of Holden's admissibility argument — that the tests did not prove parentage — arises again in our examination of whether Pioneer provided sufficient evidence to prove that Holden possessed Pioneer's protected material. This argument will be considered later, but we begin with a discussion of the specific tests and test results presented to the district court.

B.

The three tests played a significant role in evaluating Holden's claims as to how it developed the contested lines. Holden asserts that each of these lines developed from a common parent, L120. Because Holden discarded L120 two years before Pioneer filed suit when it began putting its inventory on computer and discarded a number of "old lines," the parties could not directly test L120. Instead, the parties attempted to reconstruct the genetic make-up of L120 from its pedigree supplied by Holden. Much of the scientific testimony in this case sought to prove the genetic makeup of L120. This, in turn, would indicate whether L120 could have been the parent of LH38-39-40, and whether it derived from protected Pioneer material.

Holden claims it developed L120 during repeated backcrossings between a disease resistant (Ht) seed and a public line, Oh43, which is freely available to breeders such as Holden. Holden stated that plant breeder Arlyn Eggerling selected certain plants (later labeled L120) from this backcross program. These plants were "tagged and bagged" but never entered into the nursery records. Holden explained that it could not provide a more accurate pedigree because inadequate nursery records existed as to L120's early development. Holden's L120 pedigree provided to the court placed its selection at approximately the fifth to eighth backcrossing. Other testimony suggested that the selection may have occurred even later in the backcross program, perhaps as late as the thirteenth generation. Holden's statement that it initially selected L120 from the fifth backcross generation or later proved increasingly significant as the trial progressed. The experts agreed that a plant randomly selected so late in a backcrossing program would theoretically consist almost entirely of Oh43 genetic material.

In an attempt to evaluate the parties' competing claims, the court oversaw three series of tests — electrophoresis, reverse phase high-performance liquid chromatography and growouts. Each test was supervised by the court, performed by independent experts, and monitored by the parties.

By consent and court order, isozyme electrophoresis tests were conducted. These tests rely upon the general principle that an allele found in a line of corn (e.g., LH38-39-40) must have been provided by one of the two parents. Alleles are the two individual genes in a particular gene pair. Thus, the presence of alleles in the child which are not in either of its purported parents discounts the possibility of parentage. According to Dr. Charles Stuber, the court-appointed expert who conducted the tests, electrophoresis indirectly examines these alleles by separating and then analyzing the various enzymes found in plants. This analysis of the enzymes, stated Stuber, informs the tester of a plant's genetic structure because of "one of the central dogmas of genetics" — the one-to-one correlation between genes and enzymes — which, he testified, is "essentially . . . true."

According to Stuber, LH38 contained isozyme allele 2.0 ("# 2 allele") at a particular locus, Pgd 1. At this same locus, one of its alleged parents, A619, contained a different allele. Thus, the #2 allele found in LH38 presumably came from LH38's other alleged parent — L120. Since Holden lost or discarded L120 two years before Pioneer filed its complaint, no L120 seed could be directly tested. However, Holden's theory of parentage indicates that L120 resulted from repeated backcrossings with Oh43. As referenced earlier, Drs. Stuber and Stephen Smith agreed that if L120 were randomly selected from the Ht backcross program, over 98% of its genetic material should be identical to Oh43. Relying on the fact that the 3.8 allele (found on Oh43) should have been found in L120, and hence at locus Pgd 1 of LH38, Stuber concluded that Holden's L120 story was "possible but . . . highly unlikely." Smith reached a similar result, testifying that the "isozyme data and the [L120] pedigree don't match" while Pioneer's stated pedigree was "absolutely consistent." Stuber reached a similar conclusion with respect to LH40.

Holden contends the # 2 allele resulted from W64A, the disease resistant seed in its L120 pedigree, and that Pioneer failed to show otherwise. Holden attacks the foundation for Pioneer's assertion that L120 would necessarily be 98% Oh43. Among the possibilities Holden cited were a much greater degree of contamination and relic heterozygosity in inbred corn lines than commonly assumed, or non-Mendelian events such as mutation. Holden also challenges the validity of the general principle underlying electrophoresis testing (i.e., a given allele must come from one of two parents). After considering these attacks, the district court accepted Stuber's testimony which showed that "Holden's two lines — LH38 and LH40 — might well be fathered by Pioneer's H3H, but would not be fathered in the way Holden claims they were, as the probability is so low that it is not logical." Slip op. at 34-35, 40. The duty of weighing the strength of the testimony supporting the parties' respective views lies with the district court. Hoefelman v. Conservation Comm'n, 718 F.2d 281, 285 (8th Cir. 1983); Lansford-Coaldale Joint Water Auth. v. Tonolli Corp., 4 F.3d 1209, 1218 n. 7 (3d Cir. 1993). Thus, we cannot say that the district court clearly erred in finding that L120 would consist primarily, although not entirely, of genetic material found in Oh43.

The electrophoresis evidence does not eliminate the possibility that L120 is a parent of LH38-39-40. As with such tests as typical DNA testing, see Martinez, 3 F.3d at 1193, the electrophoresis testing (even under Pioneer's interpretation) did not and could not conclusively prove that a given child (i.e., LH38-39-40) derived from particular parents. See also Bonds, 12 F.3d at 558 (DNA testing). Nonetheless, the court did find, based on Stuber's testimony, that electrophoresis testing "is very precise" and "a very important technique now being utilized by the seed corn industry on a day-to-day basis." Slip op. at 39. "All parties," the court concluded, "agreed that electrophoresis cannot recognize all differences, but that it can establish differences and that it cannot establish sameness." Id. at 32 n. 5 (emphasis in original).

The details of reverse phase high-performance liquid chromatography testing are also complex, but basically similar to electrophoresis. Like electrophoresis, liquid chromatography testing identifies certain gene proteins within the genetic structure of a plant. According to Stuber and Bietz, this process examines different genes from those measured by the electrophoresis tests. This is significant. Since the two tests measure different gene structures and are independent, their combined accuracy is higher than if either was conducted alone. Liquid chromatography testing creates a graph consisting of a series of peaks and valleys for a given seed corn. These peaks and valleys represent collections of genetic material, which theoretically must have come from one of the parents. The testimony showed that although its application to corn is relatively new, this methodology is "well established in testing other cereal grains." Slip op. at 44.

Pioneer relies on the presence of two peaks in LH39 which are not present in Oh43. Holden claims that LH39 is the result of a cross between L120 and Oh43. Pioneer responds that since L120 is "essentially Oh43 itself . . . LH39 should . . . have a chromatogram identical to Oh43." Thus, according to Pioneer, the peaks indicate that L120 could not be a parent, while supporting the theory that LH39 derived from H3H/H43SZ7. LH38 also displayed a peak inconsistent with Holden's asserted pedigree. Pioneer's position hinges on the finding that L120 would necessarily have an identical chromatogram to Oh43. If L120 differed significantly from Oh43, the existence of a couple of differing chromatogram peaks in the contested seed might very well be expected. Based on the evidence before it, the court concluded that L120 should exhibit a chromatogram identical to Oh43. Slip op. at 46.

As with electrophoresis, liquid chromatography testing could not determine the precise parentage of LH38-39-40. Rather, it could only exclude certain possibilities. As with the electrophoresis testing, the liquid chromatography testing failed to rule out H3H/H43SZ7 as a possible parent, while raising doubts as to the viability of Holden's L120 story. The court found it significant that Pioneer wanted the tests conducted, since the tests could "torpedo" Pioneer's story but not conclusively confirm it. Slip op. at 54.

At Holden's request, the court also directed three seasons of growing, cross-pollinating, and comparing the various disputed lines. The results, particularly with respect to LH39, demonstrated similar physical characteristics and comparable degrees of hybrid vigor (heterosis) among several of the plants. Pioneer's witnesses, George Sprague and Richard L. McConnell, concluded that this evidence cast doubt on Holden's asserted L120 pedigree, and that the seeds more likely derived from H3H/H43SZ7. McConnell testified that Holden's LH38 and LH39 are too similar to Pioneer's H3H to have been developed independently.

Holden interprets the growout evidence quite differently. Holden correctly states that the growout evidence reflected differences as well as similarities between the Holden and Pioneer lines. Holden points out that Pioneer's own witness, McConnell, stated that the growouts showed only the "possibility" of derivation from H3H. Dr. Paul Cornelius and Mr. Art Johnson, witnesses for Holden, looked at different traits from those analyzed by Sprague and McConnell, and found substantial differences between LH38 and H3H. Based on this investigation, Cornelius expressed his serious doubts about whether LH38-39 derived from Pioneer's line. Finally, any perceived similarities in the growouts, Holden suggests, merely reflect the large percentage of Oh43 in each of the tested lines — Holden's and Pioneer's alike.

The court rejected Holden's attempt to attack the concept or purpose of the growouts, which Holden initially requested. Slip op. at 64-65. Moreover, the court acknowledged the growouts' limitations, but concluded they were "imperfect but reliable." Id. at 65. The results, however, were "sweet and sour pork" insofar as there were "pluses and minuses for each litigant." Id. at [*]28.

Each of the three tests certainly had its own set of limitations and inadequacies. Electrophoresis and chromatography could not, even under the court's view, conclusively prove the parentage of LH38-39-40. The growouts were, as we just discussed, indecisive. Nonetheless, all of the tests contributed to the court's finding that Holden's L120 story was highly unlikely, while Pioneer's theory of parentage was more likely.

II.

Much of Holden's argument before this court focused on the district court's ruling that Holden misappropriated Pioneer's protected trade secret. According to Iowa law, a plaintiff must generally show three elements to prevail in a trade secret action: "(1) existence of a trade secret, (2) acquisition of the secret as a result of a confidential relationship, and (3) unauthorized use of a secret." Basic Chemicals, Inc. v. Benson, 251 N.W.2d 220, 226 (Iowa 1977); see Hudson Hotels Corp. v. Choice Hotels Int'l, 995 F.2d 1173, 1176 (2d Cir. 1993) (applying New York law). We review determinations of state law de novo, giving the district court no deference. Salve Regina College v. Russell, 499 U.S. 225, 231, 111 S.Ct. 1217, 1221, 113 L.Ed.2d 190 (1991).

Holden does not argue on appeal that genetic messages cannot qualify as trade secrets. Ronald Holden, Holden's chief operating officer, testified that he does not "have any problem" with the idea that there is a claimed proprietary interest in inbreds and parent lines. Thus, we assume without deciding that genetic messages can qualify for trade secret status. See Restatement of Torts, § 757 cmt. b (1939). Holden instead argues that it should not be liable for misappropriating H3H/H43SZ7 because Pioneer failed: (1) to keep the genetic messages secret; (2) to prove that Holden actually possessed the protected genetic messages; and (3) to prove that Holden obtained the material by improper means. The district court rejected each of Holden's arguments.

Insofar as Holden argues that there is insufficient evidence to support such findings, we must review the evidence in the light most favorable to Pioneer, giving it all favorable inferences. Craft v. Metromedia, Inc., 766 F.2d 1205, 1212 (8th Cir. 1985), cert. denied, 475 U.S. 1058, 106 S.Ct. 1285, 89 L.Ed.2d 592 (1986); Dace v. ACF Indus., Inc., 722 F.2d 374, 375-76 (8th Cir. 1983) (subsequent history omitted). For the most part, Holden's arguments relate to the insufficiency of the evidence or Pioneer's failure of proof. On occasion, however, Holden argues that the district court's findings are clearly erroneous. See Anderson v. City of Bessemer City, 470 U.S. 564, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985). Findings of fact are clearly erroneous only if they inspire a "definite and firm conviction that a mistake has been made." Id. at 573, 105 S.Ct. at 1511 (citation omitted). Of course, "[i]f a finding is not supported by substantial evidence, it will be found to be clearly erroneous." 9 Charles A. Wright Arthur R. Miller, Federal Practice and Procedure § 2585 (1971). But, where "the district court's account of the evidence is plausible in light of the record viewed in its entirety, the court of appeals may not reverse it even though convinced that had it been sitting as a trier of fact it would have weighed the evidence differently." Anderson, 470 U.S. at 573-74, 105 S.Ct. at 1511.

A.

Holden argues that H3H/H43SZ7 are not trade secrets because Pioneer failed to maintain their secrecy. Fundamental to the existence of a trade secret is that the matter be, in fact, secret. Kewanee Oil Co. v. Bicron Corp., 416 U.S. 470, 475, 94 S.Ct. 1879, 1883, 40 L.Ed.2d 315 (1974); Coenco, Inc. v. Coenco Sales, Inc., 940 F.2d 1176, 1178-79 (8th Cir. 1991). By definition, trade secret law does not protect information in the public domain or otherwise readily ascertainable. Kewanee Oil, 416 U.S. at 481, 484, 94 S.Ct. at 1886, 1887 ("that which is in the public domain cannot be removed therefrom by action of the States"); see Integrated Cash Mgmt. Serv., Inc. v. Digital Transactions, Inc., 920 F.2d 171, 174 (2d Cir. 1990). The secrecy, however, need not be absolute. Reasonable precautions to protect the secrecy of a trade secret will suffice. See Surgidev Corp. v. Eye Technology, Inc., 828 F.2d 452, 455 (8th Cir. 1987) ("[o]nly reasonable efforts, not all conceivable efforts, are required"); Sigma Chem. Co. v. Harris, 794 F.2d 371, 373-74 (8th Cir. 1986) (employer took extensive measures to safeguard its product and vendor files); Restatement of Torts § 757 cmt. b (1939). Companies need not "guard against the unanticipated, the undetectable, or the unpreventable methods of espionage now available" or create "an impenetrable fortress." E.I. duPont deNemours Co. v. Christopher, 431 F.2d 1012, 1016-17 (5th Cir. 1970), cert. denied, 400 U.S. 1024, 91 S.Ct. 581, 27 L.Ed.2d 637 (1971); see Rockwell Graphic Sys., Inc. v. DEV Indus., Inc., 925 F.2d 174, 180 (7th Cir. 1991) ("perfect security is not optimum security" and therefore is not required).

The district court found that the genetic messages of H3H and H43SZ7 were trade secrets. Slip op. at 72. The district court found that the "formula" did not exist outside Pioneer's and its contractors' fields, and that Pioneer took reasonable precautions to protect the secrecy of the genetic message. Id.

Pioneer takes several measures to preserve the secrecy of its inbreds. Growers operate under contracts which prohibit disclosure of the seed. Fields have no labels indicating what seed is being grown, and all seed bags are coded to avoid identification. Pioneer removes male inbred lines and commingles them with other corn, thereby frustrating those seeking to obtain the inbred seed.

Despite these steps, Holden argues that Pioneer failed to maintain adequate security in several ways. First, during the 1970s Pioneer harvested the male rows of its production fields and sold the seed to grain elevators. This grain was readily available to any purchaser. Holden also states that H3H could be present in Hybrid Pioneer 3780 or 3541 seed sold to farmers, and thus H3H plants may be found in farmers' fields. Holden further argues that Pioneer lost the requisite secrecy by selling H3H/H43SZ7 in the Soviet Union without restriction. Holden cites testimony indicating that some of this seed was "leaked out," and that Pioneer taught experts in Hungary how to produce H3H and H3H-derived hybrids.

Although there is conflicting evidence on whether Pioneer had maintained the secrecy of the genetic messages, there is sufficient evidence to support the district court's finding that Pioneer took reasonable precautions to protect the secrecy of the genetic message of H3H/H43SZ7. See Trandes Corp. v. Guy F. Atkinson Co., 996 F.2d 655, 664 (4th Cir.) ("secrecy is a question of fact"), cert. denied, ___ U.S. ___, 114 S.Ct. 443, 126 L.Ed.2d 377 (1993). First, Holden's claim that the message was "widely disseminated" because it could be obtained from seed elevators and farmers' seed bags is overstated. Although a Pioneer employee, Thomas Urban, acknowledged that the male corn could get into a seed bag as a contaminant, he also stated "it's almost impossible." Urban explained that Pioneer had "many, many employees who stood looking at corn ears . . . [and] [t]he chances of that happening are not very great." Urban knew of no instances of the male corn being found in bags of hybrid seed. Moreover, the seed in the bags and elevators is mixed, and a grower would have no way of easily identifying the desired male seed from the many other seeds in the overall mix. "The more difficult, time consuming and costly it would be to develop the information, the less likely it is `readily' ascertainable." Peter J. Courture, "Independent Derivation and Reverse Engineering," in Trade Secret Protection and Litigation: Protecting Confidential Business and Technical Information 623 (PLI Patents, Copyrights, Trademarks Literary Property Course Handbook Series No. 340, 1992) (citing cases); see Integrated Cash Mgmt. Serv., 920 F.2d at 174 (product was not "readily ascertainable"); Sigma Chem. Co., 794 F.2d at 374 (finding that vendor files constituted a trade secret and relying, in part, on fact that it would be difficult for a competitor to duplicate the information).

It is also relevant that the possible means identified by Holden of obtaining H3H/H43SZ7 were not the means by which Holden claimed it developed LH38-39-40. Many courts have held that the fact that one "could" have obtained a trade secret lawfully is not a defense if one does not actually use proper means to acquire the information. See, e.g., Bryan v. Kershaw, 366 F.2d 497, 500-01 (5th Cir. 1966), cert. denied, 386 U.S. 959, 87 S.Ct. 1030, 18 L.Ed.2d 108 (1967); Smith v. Dravo Corp., 203 F.2d 369, 374-75 (7th Cir. 1953); Imi-Tech Corp. v. Gagliani, 691 F. Supp. 214, 231 (S.D.Cal. 1986); Syntex Opthalmics, Inc. v. Novicky, 214 U.S.P.Q. 272, 277, 1982 WL 63797 (N.D.Ill. 1982) ("[w]here the security lapses were not the cause of the misappropriation, those lapses should not be the basis for denying protection"), aff'd, 701 F.2d 677 (7th Cir. 1983); Courture, "Independent Derivation and Reverse Engineering," at 623 (citing cases); but see Van Prods. Co. v. General Welding and Fabricating Co., 419 Pa. 248, 213 A.2d 769, 778-79 (1965). This evidence does not lead us to conclude that the district court clearly erred in finding that Pioneer maintained the requisite secrecy for trade secret protection of H3H/H43SZ7.

Holden's argument that Pioneer abandoned its trade secret by selling H43SZ7 to the Soviet Union is similarly unpersuasive. Pioneer did sell H43SZ7 to the Soviet Union during the 1970s. This sale, however, was pursuant to an agreement which restricted use of the seed and contained a confidentiality provision. See Trandes Corp., 996 F.2d at 664 (limited disclosure pursuant to a confidentiality agreement does not destroy secrecy). Holden's evidence that the material was "leaked out" of the Soviet Union to other Eastern bloc countries is based only on rumor, and does not convince us that the district court clearly erred in finding that Pioneer maintained the requisite secrecy for trade secret protection of H3H/H43SZ7.

B.

Holden also argues that Pioneer's trade secret claim is invalid because insufficient evidence supports the district court's finding that Holden possessed H3H or H43SZ7. Holden contends that since none of the tests can conclusively prove parentage, the district court clearly erred in finding possession. Holden points out particular shortcomings with each of the tests. First, tables provided by Stuber suggest that electrophoresis is less accurate than Pioneer's experts claimed. More importantly, Pioneer's experts' conclusions, drawn from both the electrophoresis and liquid chromatography testimony, rested on the assumption that L120 was essentially Oh43 — an assumption which Holden brought into question. Finally, Holden emphasizes that the court found the growout testimony to be indecisive.

Holden's argument rests on its claim that the scientific evidence was based on methodology not generally accepted in the scientific community. Holden's argument relies on Frye and Two Bulls which, as we have demonstrated, supra pages 1229-30, is no longer tenable in light of Daubert. Further, relying on Chaney v. Smithkline Beckman Corp., 764 F.2d 527 (8th Cir. 1985), Holden argues that the evidence was insufficient to warrant a finding that Holden derived LH38-39-40 from H3H/H43SZ7. Holden argues that the evidence showed only that it is possible that Holden's lines were derived from Pioneer's, and that to establish possession as a scientific fact, Pioneer must show more than a mere possibility of derivation.

In Chaney, we affirmed the district court's entry of summary judgment for four drug manufacturers because the plaintiff failed to present a submissible issue of causation. Id. at 529. Because plaintiff's "most favorable characterization showed only that there was a 20-80% probability" of causation, we held that such speculative testimony was insufficient to take the case to the jury. Id. Thus, the question in Chaney was the degree of certainty required by scientific evidence before a jury can consider it. Holden makes the same argument with respect to the evidence before us, essentially arguing that there was insufficient evidence to support the district court's finding that Holden derived LH38-39-40 from H3H/H43SZ7.

We believe that there is sufficient evidence to warrant a finding that Holden derived LH38-39-40 from H3H/H43SZ7. Holden does not attack the qualifications of the various experts, and our review of their testimony convinces us that the experts established by a preponderance of the evidence that Holden derived LH38-39-40 from H3H/H43SZ7. See Twin City Plaza, Inc. v. Central Sur. Ins. Corp., 409 F.2d 1195, 1203 (8th Cir. 1969). Despite Holden's claims to the contrary, this is not a case such as Chaney in which the expert testimony established only that plaintiff's theory was merely "possible" or only 20-80% likely. 764 F.2d at 529. The various tests, according to Pioneer's witnesses, indicated that Holden more likely than not used Pioneer's trade secrets to develop LH38-39-40. One Pioneer expert, Dr. Donald Duvick, estimated the likelihood of Holden's independent development of L120 to be approximately one in a trillion. Sprague stated that he found it "extremely difficult" to buy Holden's L120 story. Although reasonable minds could differ as to the merits of the experts' various conclusions, the testimony certainly permitted the court to do far more than "speculate or guess" about possession. Chaney, 764 F.2d at 529 (quoting Chicago Great Western Ry. v. Smith, 228 F.2d 180, 182 (8th Cir. 1955)). The combined results of all the tests, when viewed together, provide sufficient evidence of possession.

Moreover, although there are competing views of the evidence, we cannot conclude the district court erred in its finding of possession. Hulsenbusch v. Davidson Rubber Co., 344 F.2d 730, 734 (8th Cir. 1965) (question of possession of a trade secret is a fact reviewed under the clearly erroneous standard), cert. denied, 382 U.S. 977, 86 S.Ct. 545, 15 L.Ed.2d 468 (1966). The issue of derivation involved consideration of complex scientific principles. The record shows this was the central issue at trial and was fully explored by the parties, with conflicting opinions offered by the competing experts. See Brooktree Corp. v. Advanced Micro Devices, Inc., 977 F.2d 1555, 1573 (Fed. Cir. 1992). "[We] will not disturb the district court's decision to credit the reasonable testimony of one of two competing experts." Lansford-Coaldale Joint Water Auth. v. Tonolli Corp., 4 F.3d 1209, 1218 n. 7 (3d Cir. 1993); Hoefelman, 718 F.2d at 285.

C.

We next consider Holden's argument that Pioneer failed to prove the second requisite element of a trade secret action. Although frequently listed as such, a "confidential relationship" is not in fact a prerequisite to a trade secret action. Holden admits that a contractual or confidential relationship is not required to maintain a trade secret action. The district court held, and the parties agree, that a plaintiff may prevail in the absence of such a relationship by showing that the secret was obtained by improper means. See Restatement of Torts § 757(a) (1939) (the critical inquiry is whether the defendant obtained the secret by "improper means"). Conduct which is not otherwise unlawful may nonetheless be improper. Christopher, 431 F.2d at 1015 (use of aircraft to view partially-completed, guarded building of competitor held to be "improper means"). Holden contends, however, that Pioneer failed to prove improper means. In support of its position, Holden suggests there is absolutely no evidence of misappropriation or improper means. Thus, Holden argues we must reverse because mere possession of a trade secret cannot support a trade secret award. See Rice Researchers, Inc. v. Hiter, 512 So.2d 1259, 1269-70 (Miss. 1987); Restatement of Torts § 757, cmt. a (1939) ("[i]t is the employment of improper means to procure the trade secret, rather than the mere copying or use, which is the basis of the liability"); cf. Kewanee Oil, 416 U.S. at 476, 94 S.Ct. at 1883 ("trade secret [law] . . . does not offer protection against discovery by fair and honest means, such as by independent invention, accidental disclosure, or by so-called reverse engineering, that is by starting with the known product and working backward to divine the process which aided in its development or manufacture").

After considering Holden's arguments, the district court found "that Pioneer has met its burden regarding misappropriation." Slip op. at 74. Although the court failed to elaborate upon the precise facts which supported its finding of improper means, such elaboration is not required.

Certainly, the issue is close. Pioneer presented no direct evidence regarding how Holden obtained H3H/H43SZ7. However, "direct evidence of industrial espionage is rarely available" and not required. Greenberg v. Croydon Plastics Co., 378 F. Supp. 806, 814 (E.D.Pa. 1974), vacated in lieu of permanent injunction, 184 U.S.P.Q. 27, 378 F. Supp. 806 (E.D.Pa. 1974); see SI Handling Sys., Inc. v. Heisley, 753 F.2d 1244, 1261 (3d Cir. 1985). "Wrongful taking of a trade secret can be found based on circumstantial evidence." Roger M. Milgrim, Trade Secrets § 15.01[1], at 15-18 n. 10 (1993) (collecting cases). Moreover, the fact that Holden may have altered Pioneer's secret does not insulate it from liability. See Superior Flux Mfg. Co. v. H S Indus., Inc., 210 U.S.P.Q. 669 (N.D.Ohio 1980) (alterations to protected formula would not prevent liability). Slight alteration of a competitor's protected secret implicates the same concerns as cases of outright misappropriation.

We examine the district court's finding of misappropriation under a clearly erroneous standard. See Vekamaf Holland B.V. v. Pipe Benders, Inc., 696 F.2d 608, 612 (8th Cir. 1982); Amp Inc. v. Fleischhacker, 823 F.2d 1199, 1207 (7th Cir. 1987) (applying clearly erroneous standard to district court's finding that defendant did not prove misappropriation of trade secret).

The record displays a long history of Holden attempts to obtain Pioneer's genetic material. Roland Holden, the now-retired founder of Holden's Foundation Seeds, Inc., testified about his repeated searches for Pioneer's seed. As the court stated, "it is very clear that for a number of years, Roland Holden was doing anything he could to try to find out more about and grow Pioneer lines." Slip op. at 29. These efforts included searching "friendly farms" for stray inbred plants (e.g., H3H) — a process by which he admitted obtaining possession of several Pioneer lines. Id. at 28-30. Although the court concluded that "Pioneer has not specifically shown" that these efforts were the "exact source" of L120, the testimony supports such an inference. Id. at 29. Holden's inadequate explanation of its faulty record-keeping and the untimely disposal of all L120 seed also give rise to an inference of misappropriation. See Dillon v. Nissan Motor Co., 986 F.2d 263, 269 (8th Cir. 1993) (jury could infer from plaintiffs' failure to preserve automobile for testing that destroyed evidence was unfavorable); Rockingham Machine-Lunex Co. v. National Labor Relations Bd., 665 F.2d 303, 304 (8th Cir. 1981) ("when a party has relevant evidence within its control which it fails to produce, that failure gives rise to an inference that the evidence is unfavorable to the party"), cert. denied, 457 U.S. 1107, 102 S.Ct. 2907, 73 L.Ed.2d 1316 (1982); see generally Charles McCormick, Evidence § 185 (4th ed. 1992) (discussing negative inferences).

An inference of misappropriation from limited facts is especially warranted in situations such as this where the secret itself is so unique that any form of duplication would probably be improper. Rockwell, 925 F.2d at 178 (stating that if a plaintiff can show that the probability of the defendant obtaining the trade secret by proper means is slight, then "it will have taken a giant step toward proving what it must prove" to recover in a trade secret action); Anaconda Co. v. Metric Tool Die Co., 485 F. Supp. 410, 421-22 (E.D.Pa. 1980). Unlike the situation where a competitor is found to have possession of a process, product or fact which is similar to plaintiff's "secret," this case involves possession of a product derived from the protected secret. Proof of derivation removes the possibility of independent development, further supporting the court's finding that Holden misappropriated genetic material from Pioneer.

The court did not end its analysis after determining that Pioneer met its burden as to misappropriation or improper means. Rather, the court concluded that since Pioneer "met this burden, Holden has the burden of showing that H3H and/or H43SZ7 were lawfully acquired" — a burden which the court concluded Holden did not meet. Holden construes this language as expressing the court's decision to place upon it, and not Pioneer, the burden of proof on the issue of improper means.

There is authority supporting a shift in the burden of proof once a plaintiff proves that the defendant obtained and used the trade secret. See 2 Rudolf Callmann, The Law of Unfair Competition, Trademarks and Monopolies § 14.40 (4th ed. 1993). However, the cases employing a burden-shifting analysis typically involve breaches of confidential relationships. See, e.g., GarterBare Co. v. Munsingwear, Inc., 723 F.2d 707, 714-15 (9th Cir.), cert. denied, 469 U.S. 980, 105 S.Ct. 381, 83 L.Ed.2d 316 (1984); Schreyer v. Casco Prods. Corp., 190 F.2d 921, 922 (2d Cir. 1951), cert. denied, 342 U.S. 913, 72 S.Ct. 360, 96 L.Ed. 683 (1952). Nonetheless, a burden-shifting analysis may be appropriate in those instances, such as this case, where the defendant actively pursued plaintiff's secrets, discarded highly relevant information, consistently denied obtaining the protected material through any means, and possessed secrets shown to have probably been derived from (not merely exhibiting similarities to) plaintiff's secret.

It was only after finding sufficient evidence of "misappropriation" by Holden that the court first discussed shifting the burden of proving lawful acquisition to Holden. Slip op. at 74. The court's later discussion of "burden shifting" merely expresses its appreciation of the fact that once Pioneer produced convincing evidence of misappropriation, Holden was obligated to provide persuasive evidence of lawful derivation. We reject Holden's argument that the court improperly placed the burden on it to show that it lawfully acquired H3H/H43SZ7. The court's procedure can simply be read as an allocation of the burdens of going forward with the evidence. After weighing and balancing the parties' competing evidence, the district court found Holden's L120 story less persuasive than Pioneer's misappropriation claim, hence the district court held in favor of Pioneer. In light of the evidence discussed above, and after reviewing the entire record, we cannot say the district court's finding of misappropriation was clearly erroneous.

III.

Holden makes several additional arguments on appeal. We briefly consider these remaining claims. First, Holden argues that the district court erred in its holdings on Pioneer's Lanham Act, unjust enrichment, intentional interference with a business advantage, and conversion theories. We need not consider the merits of the alternate state law theories of recovery because we affirm the district court's trade secret award.

A.

We briefly consider Holden's argument that Pioneer's claim of reverse passing off in violation of section 43(a) of the Lanham Act is "completely without merit." 15 U.S.C. § 1125(a) (1988).

The district court concluded that Holden's misappropriation of H3H/H43SZ7 and subsequent marketing of LH38-39-40 violated section 43(a) of the Lanham Act. Slip op. at 68-70. Section 43(a) prohibits the use of false descriptions or false designations of origin in the advertising or selling of goods or services in commerce. Holden does not dispute that it sells and advertises LH38-39-40 in commerce, but argues that it has not falsely designated the origin of LH38-39-40.

The typical Lanham Act claim involves one of two factual patterns: (1) a defendant's false advertising of its goods or services; or (2) the selling or "palming off" by a defendant of its goods by use of a competitor's name. Lamothe v. Atlantic Recording Corp., 847 F.2d 1403, 1406 (9th Cir. 1988). The statute, however, extends beyond these isolated patterns, reaching those situations which are "economically equivalent to palming off." Smith v. Montoro, 648 F.2d 602, 605 (9th Cir. 1981) (internal quotation omitted). Included in this category are claims of "reverse palming off," which Pioneer alleges in this case. Id.; Williams v. Curtiss Wright Corp., 691 F.2d 168, 172 (3d Cir. 1982).

Reverse palming off is essentially the defendant's unauthorized removal of plaintiff's product's identifying marks before reselling the goods. Montoro, 648 F.2d at 605; Web Printing Controls Co. v. Oxy-Dry Corp., 906 F.2d 1202, 1203 n. 1 (7th Cir. 1990). The doctrine includes situations in which a defendant markets another's product that has been only slightly modified and then relabeled. See Roho, Inc. v. Marquis, 902 F.2d 356, 359 (5th Cir. 1990); Arrow United Indus., Inc. v. Hugh Richards, Inc., 678 F.2d 410, 412, 415 (2d Cir. 1982).

As discussed above, the district court found that Holden possessed H3H/H43SZ7 in the form of LH38-39-40, and this finding is not clearly erroneous. It is uncontested that Holden marketed LH38-39-40 as an independently-developed product. Neither its advertising nor its registration under the Plant Variety Protection Act referred to the existence of Pioneer's genetic material in its pedigree. This misrepresentation as to the origin of LH38-39-40 implicates several concerns protected by the Lanham Act's prohibition of reverse palming off. For instance, Holden's claims of independent development denied Pioneer "the advertising value of its name and of the goodwill that otherwise would stem from public knowledge of the true source of the satisfactory product." Smith, 648 F.2d at 607. "The ultimate purchaser is harmed as well by the loss of knowledge of and possible deception regarding the true source of the product or service." Roho, 902 F.2d at 359. In light of these concerns and on the facts before it, the district court did not err in concluding that Holden violated the Lanham Act by reverse palming off Pioneer's genetic material as its own.

B.

Finally, Holden argues that because Pioneer's Lanham Act claims were "completely without merit," the district court lacked federal subject matter jurisdiction over Pioneer's pendent state claims. See Levering Garrigues Co. v. Morrin, 289 U.S. 103, 105, 53 S.Ct. 549, 550, 77 L.Ed. 1062 (1933) (no subject matter jurisdiction where federal claim is "plainly unsubstantial"). As we have affirmed the judgment in favor of Pioneer on its Section 43(a) claim, we reject Holden's argument.

Even if we were to have reversed the Lanham Act judgment, however, we would still reject Holden's jurisdictional claim. The federal courts do have a duty to examine the substantiality of the federal claim throughout the litigation, and must dismiss all claims if the federal claim proves patently meritless after the trial begins. United Mine Workers v. Gibbs, 383 U.S. 715, 725, 727, 86 S.Ct. 1130, 1138, 1139-40, 16 L.Ed.2d 218 (1966) (question of substantiality "remains open throughout the litigation"); see Tully v. Mott Supermarkets, Inc., 540 F.2d 187, 195-96 (3d Cir. 1976). However, pendent jurisdiction is a discretionary doctrine rooted "in considerations of judicial economy, convenience and fairness to litigants." Gibbs, 383 U.S. at 726, 86 S.Ct. at 1139. It is the law in this circuit that "the substantial investment of judicial time and resources in the case . . . justifies the exercise of jurisdiction over the state claim, even after the federal claim has been dismissed." North Dakota v. Merchant Nat'l Bank Trust Co., 634 F.2d 368, 371 (8th Cir. 1980) (en banc); see Gilbert/Robinson, Inc. v. Carrie Beverage — Missouri, Inc., 989 F.2d 985, 993 (8th Cir.), cert. denied, ___ U.S. ___, 114 S.Ct. 338, 126 L.Ed.2d 282 (1993); First Nat'l Bank Trust Co. v. Hollingsworth, 931 F.2d 1295, 1307-08 (8th Cir. 1991).

With these considerations of judicial economy in mind, we would reject Holden's argument that this action must be dismissed and brought before a state court, even if we had reversed the Lanham Act award. This case, after all, consumed the attention of the district court for over a decade, required ten weeks of actual trial time, and involved court-ordered complex scientific testing and growouts. Subject matter jurisdiction should not be used to dismiss a case containing even a remotely plausible federal claim if the parties and the courts have already made the vast expenditure of resources involved in this case. We entertain no doubt that Pioneer's Lanham Act claim is sufficiently meritorious to confer federal jurisdiction. Having obtained jurisdiction over this claim, the district court properly exercised its "largely unrestricted" power to entertain pendent jurisdiction. Tully, 540 F.2d at 196; see Curtis v. Sears, Roebuck Co., 754 F.2d 781, 786-87 (8th Cir. 1985).

C.

Holden also argues the Plant Variety Protection Act, 7 U.S.C. § 2321-2582 preempts state trade secret law as applied to sexually reproducing plants. Holden contends the available legislative history demonstrates Congress' intent that the Act be the sole legal protection for sexually reproducing plants. If clearly expressed, congressional intent to preempt state law will be given effect. Wisconsin Pub. Intervenor v. Mortier, 501 U.S. 597, 603-05, 111 S.Ct. 2476, 2481-82, 115 L.Ed.2d 532 (1991) (discussing four grounds for finding preemptive effect). However, neither the plain language of the Act nor the legislative history cited by Holden demonstrates any such intent. We read the House Report as reflecting Congress' determination that no patent or patent-like protection existed for sexually reproducing plants. The language simply does not show a legislative intent to preempt all state protection of plants. Congress merely acknowledged and addressed the absence of protection for sexually reproducing plants qua plants. It did not suggest that state law could not address improper conduct that happened to involve plants or plant material. The Supreme Court has expressly held that trade secret and patent protection can "peacefully coexist." Kewanee Oil, 416 U.S. at 485-87, 94 S.Ct. at 1888-89. Moreover, this court should not lightly infer preemption. International Paper Co. v. Ouellette, 479 U.S. 481, 491, 107 S.Ct. 805, 811, 93 L.Ed.2d 883 (1987); see Florida Lime Avocado Growers, Inc., v. Paul, 373 U.S. 132, 147-50, 83 S.Ct. 1210, 1220-21, 10 L.Ed.2d 248 (1963). We have carefully considered Holden's preemption argument, and conclude it is without merit.

IV.

The district court bifurcated the liability and damage trials. After considering the various submissions of the parties in the separate damage trial, the court awarded Pioneer $46,703,230 in lost profits. The court further determined that, in the alternative, the proper damages under an unjust enrichment theory would be $21,174,913. The court also enjoined Holden from distributing or disposing of 177 inbred lines containing LH38-39-40 or their progeny, pending a determination of whether such lines should be turned over to Iowa State University.

The parties' dispute over damages involves three major issues: (1) Did the district court properly choose the "lost profits" theory, instead of an unjust enrichment or reasonable royalty measure? (2) If lost profits are the appropriate measure, did the district court use an acceptable methodology in computing the lost profits? (3) Did the district court err in denying Pioneer's request for prejudgment interest?

A.

Courts have used a wide variety of methods to measure damages in trade secret cases, including lost profits, unjust enrichment and reasonable royalty. See Curtiss-Wright Corp. v. Edel-Brown Tool Die Co., 381 Mass. 1, 407 N.E.2d 319, 327 (1980); see generally Michael A. Rosenhouse, Annotation, Proper Measure and Elements of Damages for Misappropriation of Trade Secret, 11 A.L.R.4th 12 (1982 Supp. 1983); 2 Rudolph Callmann, The Law of Unfair Competition, Trademarks and Monopolies § 14.45 (1993 Supp.). Of these, the reasonable royalty theory is least plausible on our facts. Although occasionally applied in trade secret cases, see 11 A.L.R.4th § 33[a], at 94-100 (collecting cases), this theory is most appropriate when the other theories would result in no recovery or when the parties actually had or contemplated a royalty arrangement. See Curtiss-Wright, 407 N.E.2d at 327 n. 9; see generally 11 A.L.R.4th § 2[a], at 21. The district court's assessment of over 40 million dollars for lost profits and 20 million dollars for unjust enrichment suggests that this is not a case where a royalty provides the only possible basis for recovery. Nor is there any evidence that Pioneer ever licensed or considered licensing its trade secrets to any direct competitor or foundation seed company.

In general, both lost profits and unjust enrichment theories have been widely accepted. See 2 Callmann § 14.45 n. 5, at 307-08 (Cum.Supp. 1993) (collecting cases); 11 A.L.R.4th §§ 4, 11-12; Restatement of Torts § 757(e) (1939). Courts and commentators alike have suggested that in choosing among competing theories, courts typically select the measure "which affords the plaintiff the greatest recovery." 11 A.L.R.4th § 2[a], at 21; see Curtiss-Wright, 407 N.E.2d at 327; 2 Callmann § 14.45, at 170.

The district court determined that Pioneer's lost profits were the appropriate measure for damages. The Iowa Supreme Court has recognized lost profits as an appropriate measure of damages. Basic Chemicals, 251 N.W.2d at 233. The selection of lost profits as the appropriate measure in this case, however, presented formidable problems stemming from the fact that Holden does not directly compete with Pioneer. Rather, Holden provides inbred lines which its customers use to develop hybrids which directly compete with Pioneer's hybrids.

In choosing a particular method to quantify Pioneer's lost profits, the district court adopted the basic methodology of Pioneer's expert, Michael Wagner. Wagner sought to quantify the amount Pioneer would have made "but for" Holden's misappropriation. Wagner began by considering the sales of "look-alikes" from 1979 to 1989. Then, he assumed that, absent Holden's distribution of LH38-39-40, Pioneer would have obtained the same percentage of these sales as its market share in all other sales (36% average). From this, Wagner computed lost profits of $140,109,691. The court, however, reduced this amount by two-thirds reflecting its determination that Pioneer would not have obtained the full 36% market penetration in that portion of the market occupied by the "look-alike."

Holden attacks the court's analysis at several points. First, it claims the district court lacked an adequate basis for its assessment of damages. Although Holden blurs the two, there are two distinct steps to such a damage inquiry. Basic Chemicals provides the governing test:

If it is speculative and uncertain whether damages have been sustained, recovery is denied. If the uncertainty lies only in the amount of damages, recovery may be had if there is proof of a reasonable basis from which the amount can be inferred or approximated.

251 N.W.2d at 233 (quoting Northrup v. Miles Homes, Inc., 204 N.W.2d 850, 857 (Iowa 1973)). Thus, under Iowa law, a court must first determine if damage occurred, and then consider whether there is a reasonable basis for awarding a specific damage amount.

The district court found that Pioneer lost profits, and that Holden's expert had so conceded. This finding is amply supported by the record. Hybrids developed from LH38-39-40 were used for several years in direct competition with Pioneer. As demonstrated by the growouts, some of these hybrids were quite similar to Pioneer's hybrids. The availability of these hybrids (which Holden contends are actually superior to those of Pioneer) resulted in fewer Pioneer sales at perhaps lower prices. Thus, the only relevant question is the second part of the Basic Chemicals analysis — whether there was a "reasonable basis" from which the court could award a particular amount. Id.

We believe the district court had a reasonable basis for its award of $46,703,230. The court relied on Holden's actual sales figures, the known productive capacity of Pioneer's parent lines, Pioneer's profitability history and a reasonable estimate of Pioneer's lost share of the "look-alike" market. This meets the "reasonable basis" requirement of Basic Chemicals.

Holden also contends that the entire "but for" rationale is misguided. However, the court's rationale appears to be a straightforward and not uncommon approach to determining lost profits. See King Instrument Corp. v. Otari Corp., 767 F.2d 853, 863 (Fed. Cir. 1985) (lost profits award "requires (1) a showing that the patent owner would have made the sale but for the infringement, i.e., causation existed, and (2) proper evidence of the computation on the loss of profits"), cert. denied, 475 U.S. 1016, 106 S.Ct. 1197, 89 L.Ed.2d 312 (1986); but see Panduit Corp. v. Stahlin Bros. Fibre Works, 575 F.2d 1152, 1156 (6th Cir. 1978) (requiring patent owner seeking lost profits to show: (1) demand for product, (2) no non-infringing substitutes, (3) ability to exploit the demand, and (4) amount of profit it would receive). Holden's attack on this point, based on Panduit, seems largely over the verbiage the court should have employed. The court's result, we believe, would be the same under any lost profits verbiage. Accordingly, we find no error in the court's lost profits methodology and conclude the district court did not err in awarding Pioneer $46,703,230.

B.

Pioneer also requested prejudgment interest totalling $29,487,632. Relying on Stroh Container Company v. Delphi Industries, Inc., 783 F.2d 743, 751-52 (8th Cir.), cert. denied, 476 U.S. 1141, 106 S.Ct. 2249, 90 L.Ed.2d 695 (1986), the district court denied the request, finding "unusual circumstances which would make such an award inequitable," and remarking particularly that Holden could not have calculated damages so as to arrive at an early settlement offer in this case. Pioneer Hi-Bred Int'l v. Holden Found. Seeds, No. 81-60-E, slip op. at 7, 9, 1991 WL 628674 (S.D.Iowa Dec. 30, 1991). In a later ruling, the court also held that Pioneer is not entitled to prejudgment interest under Iowa law, pointing out that while Iowa Code § 535.3 (1987) is mandatory in its terms and has been so described by Iowa courts, at least five recognized exceptions exist. Pioneer Hi-Bred Int'l, Inc. v. Holden Found. Seeds, No. 81-60-E, slip op. at 3, 1992 WL 674084 (S.D.Iowa Sept. 23, 1992). The district court, without fully articulating its position, held that the equitable considerations, including the lengthy delay that could not be attributed to Holden and the substantial size of the liability award, were unusual circumstances that the court could consider in denying prejudgment interest under Iowa law. Slip op. at 7-10 (Dec. 30, 1991). The court predicted that the Iowa Supreme Court, if called upon to consider the issue, would similarly decide that Pioneer's request should be denied. Slip op. at 3 (Sept. 23, 1992). Pioneer cross-appeals from the court's denial of prejudgment interest.

The parties dispute whether the court should have applied state law in determining if prejudgment interest was warranted. Holden contends the court below correctly applied federal common law principles in denying Pioneer's request. Under federal law, interest may be withheld if "there are exceptional circumstances to justify the refusal." Cargill, Inc. v. Taylor Towing Serv., Inc., 642 F.2d 239, 242 (8th Cir. 1981). More specifically, prejudgment interest may be denied in Lanham Act cases. Neva, Inc. v. Christian Duplications Int'l, Inc., 743 F. Supp. 1533, 1542-43 (M.D.Fla. 1990). Moreover, at least one circuit has held that when faced with a general verdict resulting from mixed federal and state claims, the district court may apply the federal rule and deny interest. Wojtkowski v. Cade, 725 F.2d 127, 129 (1st Cir. 1984); but see Doty v. Sewall, 908 F.2d 1053, 1063 (1st Cir. 1990) (requiring prejudgment interest where state and federal claims are identical and state law imposes mandatory interest).

In light of the particular facts before it, the district court committed no error in refusing to award prejudgment interest to Pioneer. To the extent that federal law governs, the considerations detailed by the court, see slip op. at 6-10 (Dec. 30, 1991), adequately support its decision to exercise its equitable authority and refuse Pioneer's request. See Stroh Container Co., 783 F.2d at 752. One of the major goals thought to be served by prejudgment interest — promoting settlement — is far less compelling in a case such as this where the issue of liability is highly challenged and the damages are quite large but virtually impossible to ascertain prior to trial. See id. (award of prejudgment interest serves to "promote settlement and deter attempts to benefit unfairly from the inherent delays of litigation").

We are well aware that Iowa law expressly provides that prejudgment interest "shall be allowed on all money due on judgments . . . from the date of the commencement of the action." Iowa Code § 535.3 (1987); Iowa State Commerce Comm'n v. Manilla Grain Terminal, Inc., 362 N.W.2d 562, 565 (Iowa 1985). This language is "in general . . . mandatory." Barske v. Rockwell Int'l Corp., 514 N.W.2d 917, 925 (Iowa 1994); see Mermigis v. Servicemaster Indus., Inc., 437 N.W.2d 242, 248 (Iowa 1989); Oskaloosa Food Prods. Corp. v. Aetna Casualty Sur., 337 N.W.2d 521, 523 (Iowa 1983); Adams v. Johnson, 445 N.W.2d 422, 424 (Iowa App. 1989). However, when faced with particular situations in which an award of prejudgment interest serves little purpose or would otherwise prove inequitable, the Iowa courts have on several occasions created exceptions to this general rule. Recognized categories of exceptions include: (1) punitive damage awards: (2) damages for post-filing injuries; (3) specific performance orders; (4) future damages; (5) Iowa Tort Claims Act liability; and (6) divorce decrees. In re Marriage of Baculis, 430 N.W.2d 399, 402-04 (Iowa 1988). None of these exceptions would bar recovery in this case.

However, like the district court, we believe that the Iowa Supreme Court would create another exception if presented with the unique circumstances of this case, particularly the inherent length of the proceedings (admittedly beyond Holden's control), the substantial amount of the award and the inability to determine at the outset even the roughest estimate of damages in furtherance of settlement. Thus, we reject Pioneer's cross-appeal from the district court's denial of prejudgment interest.

We affirm the district court's judgment and order of damages.